Application of RBFN Using LPC of PD Pulse Shapes for Discriminating Among Multi PD Sources

  • Lee, Kang-Won (Safety performance Certification & Research Center, Korea Railroad Research Institute) ;
  • Lim, Kee-Joe (Dept. of Electrical Engineering, Graduate School, Chungbuk National University) ;
  • Kang, Seong-Hwa (Dept. of Industry Safety, Chungcheong College)
  • 발행 : 2003.10.01

초록

Partial discharge pulse shapes from variable PD (partial discharge) sources sustain many characteristics such as types of PD. Ultra high frequency antennas have wide bandwidth from 30KHz to 2㎓. Therefore, signals taken from a UHF antenna have important attributes (rising time, falling time, shape factor, etc.) for electromagnetic sources, such as PD sources. We investigated PD pulse shapes from several PD sources using a UHF antenna and the results were used for classification of PD sources. Features for discrimination are extracted from frequency distribution and LPC (Linear Prediction Coefficient) of time signal. RBFN are used for investigating the possibility of classification of multi-PD sources.

키워드

참고문헌

  1. R. A. Fouracre, S. J. MacGregor, F. A. Tuema, 'An Investigation into the Mechanism of Surface Discharges', 12th IEEE Intern Pulsed Power Conf , Vol. 2, pp. 1380-1383, 1999
  2. D. M. Hepburn, I. J. Kemp, A. Nesbitt, B. G. Stewart, 'Detection and characterization of partial discharge activity on outdoor high voltage insulating structures by RF antenna measurement techniques' 17th IEEE International Symposium on High Voltage Engineering, Vol. 5, pp. 188-191, 1999
  3. H.N Bidhendi, Qi Su, 'Partial discharge location on power cables using linear prediction' Proceedings of the 5th IEEE International Conference on Properties and Applications of Dielectric Materials, Vol. 1, pp. 406-409, 25-30 May 1997
  4. J. S. R. Jang, C. T. Sun, 'Neuro-Fuzzy and Soft Computting' pp. 238-241, Prentice Hall International, Inc 1977